1. Field of the Invention
The present invention relates to an image processing apparatus for processing an externally input image, image processing method and recording medium. And, more particularly, the present invention relates to an apparatus for vectorizing an illustrative image.
2. Description of the Related Art
Conventionally, a vectorization technique for fonts (character type faces) used in a wordprocessor or the like has prevailed. The conventional vectorization technique for fonts (character type faces) targets at inputs of characters which are beautifully designed in advance. This technique digitizes one character created in an analog manner to a binary image having a relatively large size of 512×512 pixels, 1024×1024 pixels, or the like, and vectorizes the edge or contour of the character based on the digitized data. Basically, since the data which is vectorized in this way allows smooth edge expression if it is enlarged to various desired sizes, it has a feature of high shape quality. Since characters of various sizes can be generated based on single data, vectorization of data can improve convenience and can reduce the data volume.
Therefore, as disclosed in Japanese Patent No. 3049672, the vectorization technique has been widely proposed not only for fonts (character type faces) but also for binary images, thus obtaining the aforementioned effects.
In recent years, as described in Japanese Patent No. 2845107, a vectorization technique of a scanned image including a full-color part has also been proposed.
Vectorization of binary images is also described in Japanese Patent No. 3026592. As described in this reference, the vectorization method functions as follows. That is, a predetermined position is set as a point that forms a contour line based on the states of a pixel of interest and its adjacent pixels, and a connection direction of points that form the contour line is determined based on the states of adjacent pixels. Furthermore, a connection state between a point that forms the contour line and another point that forms the contour line is checked, the position of a pixel of interest is updated on image data in a raster scan order, and contour points are extracted based on the states of adjacent pixels for each pixel of interest. In this arrangement, the states of a pixel of interest and its adjacent pixels in image data are held, the pixel of interest is extracted in the raster scan order, and inter-pixel vectors in the horizontal and vertical directions are detected based on the states of the pixel of interest and adjacent pixels. The connection state of the inter-pixel vectors is determined, and the contour of image data is extracted based on the determined connection state of the inter-pixel vectors. The method described in this reference can extract all contour lines in an image by only one raster scan order, and does not require any image memory used to store all image data. As a result, the memory capacity can be reduced, and the contours can be extracted from borders of pixels in place of the central positions of pixels of an input image, thus allowing extraction of a contour line having a significant width even for a thin line having a one-pixel width.
As described in Japanese Patent Laid-Open No. 2005-346137, by approximating contour information of a binary image by not only lines but also quadratic or cubic Bezier curves, the contour information that expresses a variable-scale image with high image quality can be functionally approximated to have a smaller data size.
However, in the aforementioned conventional image vectorization technique, upon storing vectorized data, the data are stored in turn top down of an image. Therefore, when only data indicating features of an object as a target are extracted from high-precision image data to render a sketch of the entire image, the processing becomes complicated.